mediocregopher/mediocre-blog
1
0
Fork 0
Code Pull Requests Releases Activity

updated generations

Browse Source
This commit is contained in:
Brian Picciano 2013-11-18 22:35:32 -05:00
parent f2025fd0ea
commit 551f66a639
1 changed files with 13 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
_posts/2013-10-8-generations.md
View File

@ -8,15 +8,15 @@ file-services.
# The problem # The problem
Working at a shop where we have millions of different files, any of which could At [cryptic.io][cryptic] we plan on having millions of different
be arbitrarily chosen to serve to a file at any given time. These files are files, any of which could be arbitrarily chosen to be served any given time.
uploaded by users of the app and retrieved by others. These files are uploaded by users at arbitrary times.
Scaling such a system is no easy task. The chosen solution involves shuffling Scaling such a system is no easy task. The solution I've seen implemented in the
files around on a nearly constant basis, making sure that files which are more past involves shuffling files around on a nearly constant basis, making sure
"popular" are on fast drives, while at the same time making sure that no drives that files which are more "popular" are on fast drives, while at the same time
are at capicty and at the same time that all files, even newly uploaded ones, making sure that no drives are at capicty and at the same time that all files,
are stored redundantly. even newly uploaded ones, are stored redundantly.
The problem with this solution is one of coordination. At any given moment the The problem with this solution is one of coordination. At any given moment the
app needs to be able to "find" a file so it can give the client a link to app needs to be able to "find" a file so it can give the client a link to
@ -56,7 +56,9 @@ starting a project from scratch
* Minimal information about where a file "is" needs to be stored. When a file is * Minimal information about where a file "is" needs to be stored. When a file is
uploaded all that's needed is to know what generation it is in, and then what uploaded all that's needed is to know what generation it is in, and then what
nodes/drives are in that generation. nodes/drives are in that generation. If the file's name is generated
server-side, then the file's generation could be *part* of its name, making
lookup even faster.
* Drives don't need to "know" about each other. What I mean by this is that * Drives don't need to "know" about each other. What I mean by this is that
whatever is running as the receive point for file-uploads on each drive doesn't whatever is running as the receive point for file-uploads on each drive doesn't
@ -86,7 +88,7 @@ from (given you have easy access to information about specific drives).
The big caveat here is that this is just an idea. It has NOT been tested in The big caveat here is that this is just an idea. It has NOT been tested in
production. But we have enough faith in it that we're going to give it a shot at production. But we have enough faith in it that we're going to give it a shot at
cryptic.io. I'll keep this page updated. [cryptic.io][cryptic]. I'll keep this page updated.
The second caveat is that this scheme does not inherently support caching. If a The second caveat is that this scheme does not inherently support caching. If a
file suddenly becomes super popular the world over your hard-disks might not be file suddenly becomes super popular the world over your hard-disks might not be
@ -94,5 +96,6 @@ able to keep up, and it's probably not feasible to have an FIO drive in *every*
generation. I think that [groupcache][1] may be the answer to this problem, generation. I think that [groupcache][1] may be the answer to this problem,
assuming your files are reasonably small, but again I haven't tested it yet. assuming your files are reasonably small, but again I haven't tested it yet.
[cryptic]: https://cryptic.io
[0]: https://github.com/cryptic-io/marlin [0]: https://github.com/cryptic-io/marlin
[1]: https://github.com/golang/groupcache [1]: https://github.com/golang/groupcache